BrombergLDEERConf2003

8/4/2019 BrombergLDEERConf2003

1/7

1

Diesel Engine Emission Reduction (DEER) Workshop, 2003

Newport RI, August 24-28, 2003

HYDROGEN GENERATION FROM PLASMATRON REFORMERS: A PROMISING TECHNOLOGYFOR NOX ADSORBER REGENERATION AND OTHER AUTOMOTIVE APPLICATIONS

L. Bromberg, MIT PSFC S. Crane, ArvinMeritor

A. Rabinovich, MIT PSFC Y. Kong, ArvinMeritor D.R. Cohn, MIT PSFC

J. Heywood, MIT Mech. Eng.

Department

N. Alexeev, Baikov Inst

Metallurgy, Moscow Ru

A. Samokhin, Baikov Inst

Metallurgy, Moscow Ru

ABSTRACTPlasmatron reformers are being developed at MIT and

ArvinMeritor [1]. In these reformers a special low power

electrical discharge is used to promote partial oxidation

conversion of hydrocarbon fuels into hydrogen and CO. The

partial oxidation reaction of this very fuel rich mixture is

difficult to initiate. The plasmatron provides continuous

enhanced volume initiation. To minimize electrode erosion and

electrical power requirements, a low current, high voltage

discharge with wide area electrodes is used. The reformers

operate at or slightly above atmospheric pressure.

Plasmatron reformers provide the advantages of rapid startup

and transient response; efficient conversion of the fuel to

hydrogen rich gas; compact size; relaxation or elimination of

reformer catalyst requirements; and capability to process

difficult to reform fuels, such as diesel and bio-oils. These

advantages facilitate use of onboard hydrogen-generation

technology for diesel exhaust after-treatment. Plasma-enhanced

reformer technology can provide substantial conversion even

without the use of a catalyst.

Recent progress includes a substantial decrease in electrical

power consumption (to about 200 W), increased flow rate(above 1 g/s of diesel fuel corresponding to approximately 40

kW of chemical energy), soot suppression and improvements in

other operational features..

Plasmatron reformer technology has been evaluated for

regeneration of NOx adsorber after-treatment systems. AtArvinMeritor tests were performed on a dual-leg NOx adsorber

system using a Cummins 8.3L diesel engine both in a test cell

and on a vehicle. A NOx adsorber system was tested using the

plasmatron reformer as a regenerator and without the reformer

(i.e., with straight diesel fuel based regeneration as the baseline

case. The plasmatron reformer was shown to improve NOregeneration significantly compared to the baseline diesel case

The net result of these initial tests was a significant decrease in

fuel penalty, roughly 50% at moderate adsorber temperatures

This fuel penalty improvement is accompanied by a dramatic

drop in slipped hydrocarbon emissions, which decreased by

90% or more. Significant advantages are demonstrated across a

wide range of engine conditions and temperatures. The study

also indicated the potential to regenerate NOx adsorbers at lowtemperatures where diesel fuel based regeneration is not

effective, such as those typical of idle conditions. Two vehicles

a bus and a light duty truck, have been equipped for plasmatron

reformer NOx adsorber regeneration tests.

INTRODUCTIONIn order to meet stringent US emissions regulations for

2007-2010 diesel vehicle model years, after-treatmen

technology is being developed. Both particulate and NO

emissions after-treatment technology will be necessary, as i

seems that engine in-cylinder techniques alone will be unable to

meet the regulations.

The purpose of the present work is to develop hydrogen

based technology for addressing after-treatment issues. In thi

paper recent developments in both the technology for on-board

generation of hydrogen as well as applications to NOx after

treatment issues will be discussed.

Potential on-board generation of hydrogen from renewable

energy biofuels for the transportation sector is also discussed


2/7

2

This application could provide a means to alleviate oil

dependency and reduce greenhouse gas emissions.

PLASMATRON TECHNOLOGYPlasma technology can be used to accelerate chemical

reactions. Two modes of operation have been explored for

promoting fuel reformation into synthesis gas (hydrogen and

carbon monoxide). The first uses thermal plasmas, where

electrons, ions and neutral particles are generally in thermal

equilibrium, and at high average temperature. This technology

has the disadvantages of high current operation (associated with

high electrode wear), and high energy consumption. Thesecond mode of operation, using special low current plasmas,

has the advantage of low power consumption and long

electrode life. However, successful operation of the plasma

process is more sensitive to design and operation. In this

section, operation of a low current plasmatron as a device for

converting hard to reform fuels to synthesis gas is described.

Fuel conversion is accomplished using the process of

partial oxidation. In this mode, enough oxygen is supplied in

order to capture each carbon atom in the fuel as carbon

monoxide, with the hydrogen released as hydrogen molecules.

The nitrogen from the air (as air is used as the oxidizer), is inertat the temperatures and pressures of operation, and thus goes

unaffected in the reactor.

The use of the plasma allows for a robust reaction initiation

for air/fuel mixture. The advantages of this technology include

fast start up and rapid response, relaxation or elimination of catalyst requirements effective reforming in compact devices, attractive for

onboard applications

efficient conversion from liquid hydrocarbons tosynthesis gas

The technology can be used for a broad range of fuels,

from light, low viscosity liquids, such as ethanol, to high

viscosity liquids, such as diesel and bio-oils.

The small size enables the use of fuel reformers for

multiple applications on vehicles. These applications include:

Diesel engine exhaust after-treatment Use of renewable energy bio-oils Fuel supply and ignition control in HCCI

(homogeneous charge compression ignition) engines

Spark ignition engines using hydrogen enhanced ultra-lean turbo-charged operation

For diesel engine exhaust after-treatment, onboard

hydrogen production can be used in the regeneration of either

NOx or particulate traps. The NOx trap regeneration application

will be discussed later in this paper. In the HCCI application,

the use of fuel reformation to supply a portion of the fuelallows adjustment of engine charge temperature and fuel

ignition characteristics providing a means of ignition control. In

the spark ignition engine application, hydrogen rich gas

addition allows ultra lean operation and can facilitate the use of

high compression ratio and strong boosting. Hydrogen enriched

combustion in a turbo-charged high compression ratio engine

can provide both increased efficiency and reduced emissions

System optimization studies indicate that a substantial increase

in fuel efficiency (up to a 30% increase in net efficiency) anddecrease in NOx emissions are achievable.

Present plasmatron reformers have volumes on the order of

1-2 liters, with an electrical power consumption of about 250

W. Figure 1 shows a photograph of a recently developed

plasmatron. The plasmatron is followed by a reaction section

which provides time for the development of homogeneousreactions. When high hydrogen yields are desired, a catalys

can be placed downstream from the homogeneous zone.

Figure 1. Compact low power plasmatron fuel converter

At present, goals for the diesel plasmatron reformer

development includes increased flow rates, operational and

configuration optimization and increased simplicity, among

other goals.

BIOFUEL REFORMATIONAt MIT, plasmatron reforming technology has been used

with for a wide range of biofuels. These fuels include corn

canola and soybean oils (both refined and unrefined), andethanol.

Reformation was achieved with a plasmatron fuel reformer

optimized for diesel fuel operation. The device operated at 200

W of electrical power, and at stoichiometry for partial oxidation

(O/C ratio ~ 1.1). The available energy from the reformate flow

rate was about 20 kW (LHV of the reformate gas, as calculated

from the LHV of compounds measured in the gas analysis)


3/7

3

Some of the oxygen required for the reformation is provided by

the fuel itself; the O/C ratio includes this oxygen,

Figure 2 shows results from the reformation as a function

of the O/C ratio, for corn and soybean oils. The hydrogen yield

is defined as the ratio of the flow rate of hydrogen in the

reformate to the flow rate of hydrogen in the fuel. A hydrogenyield of 1 thus means that all the hydrogen in the fuel has been

released as hydrogen in the reformate. High hydrogen yields

are possible (greater than 80%), even at relatively high O/C

ratios.

Figure 2. Hydrogen yield as a function of the O/C ratio,

for corn and soybean oils

The results in Figure 2 were carried out with a plasmatron

reformer that incorporated a nickel catalyst on a crushed

alumina substrate. In the absence of a catalyst (for a

homogeneous reaction), the hydrogen yields are about 40%,coupled with high concentration of light hydrocarbons (~4-5%

C2H4)

Figure 3. Hydrogen concentration in dry gas for ethanolreforming, as a function of flow rates, for O/C ~ 1.2

The presence of soot was determined using an opacity

meter. Virtually no soot was observed in these experiments.

Similar experiments for ethanol are shown in Figure 3, for

different powers and varying flow rates. Ethanol conversion is

difficult to perform homogeneously because of the low

exothermicity of the partial oxidation reaction. Figure 3 shows

the hydrogen concentration of the partial oxidation of ethanol

for O/C ~ 1.2.

DIESEL REFORMATIONThe plasmatron fuel converter has been further developed

for use with diesel fuel. The goal of reformation of diesel by

the plasmatron fuel converter is the conversion of the heavy

diesel compounds into hydrogen, carbon monoxide, and ligh

hydrocarbons for use in after-treatment applications. The diese

reformate produced by the plasmatron fuel converter has littleresidual oxygen and little or no soot. The diesel plasmatron can

be cycled on and off quickly( ~ 3-5 sec). For diesel after

treatment applications the dynamic performance of the

plasmatron fuel reformer is important because of the low duty

cycle of operation. For NOx catalyst regeneration applications

the plasmatron may be turned on for a few seconds every half

minute or so. For diesel particulate filter applications, theplasmatron could well be operated for a few minutes every few

hours resulting in much smaller duty cycles.

Results for diesel conversion without a catalyst are shown

in Table 1. Typical O/C ratios are 1.1, with opacity readings of0.0% (below the resolution of the device), for flow rates on the

order of 1 g/s. The energy conversion efficiency, defined as thelower heating value of the reformate divided by the lower

heating value of the fuel, is about 70%. The lower heating value

of the reformate is calculated using the measured composition

of the reformate.

Table I. Diesel reforming without a catalyst

If higher yields of hydrogen are desired, a catalyst can be

placed downstream from the homogeneous zone. The absenceof oxygen in the reformate due to the partial oxidation process

minimizes the occurrence of hot spots within the catalyst. This

alleviates one of the catalyst durability issues.

Hydrogen is a powerful reductant, and its use for the

regeneration of NOx traps has been researched in catalystreactor labs. The goal of the work presented next is to

determine the advantages of H2 assisted NOx trap regeneration

Electric power W 250

O/C 1.1

Diesel flow rate g/s 0.8

Corresponding chemical power kW 35

Concentration (vol %)

H2 8.2

O2 1.4

N2 68.7

CH4 2.6

CO 14.3

CO2 4.7

C2H4 2.4

C2H2 0.0

Energy efficiency to hydrogen, CO and light HC 70%

Soot (opacity meter) 0

0

20

40

60

80

100

1.0 1.2 1.4 1.6 1.8

ratio O/C

H2yield,

%

orn oil

Soybean oil

20.0

22.0

24.0

26.0

28.0

0.00 0.20 0.40 0.60 0.80 1.00

Ethanol flow rate, g/s

H2concentration,

%v

ol.

Plasma 270W

Plasma 120W


4/7


5/7

5

valve is ideal. Flow bench tests of the switching valve and

adsorber system indicated a 3% leakage rate. For the Cummins

ISC 8.3L engine used in this testing, the available plasmatron

reformate flow rate was marginal at a high engine exhaust

flows with 3% leakage. An exhaust brake valve was added to

each leg as indicated in the schematic. When the brake valve

was closed on the regenerating leg the exhaust flow leakagewas reduced to an acceptable 1%.

A photograph of the actual test cell setup at Analytical

Engineering, Inc. (AEI) is shown in Figure 6. Extensive

instrumentation was utilized including a chemi-luminescent

NOx analyzer, NOx sensors, multiple thermocouples, and a

V&F hydrogen analyzer. The fuel reformer, switching valve,and brake valves were all computer controlled. A MY2000

Cummins 8.3L engine was used in these tests. Zero sulfur

diesel fuel was used in this testing to avoid sulfur poisoning of

the traps. Also the engine oil used was a special low sulfur

formulation for the same reason. Even with these measures

some minor adsorption degradation was observed over the

course of testing.

Figure 7. Bus Road Load vs ESC 13 mode

Figure 8. NOx adsorption comparison. Bus Road Load, at

same fuel penalty (reformate vs diesel)

Two sets of engine operating points were used in this

testing: the European Stationary Cycle (ESC) 13 modes and a

calculated Bus Road Load curve. Figure 7 compares the

operating points. The ESC 13 Mode test is a steady state mode

weighted substitute for transient emissions testing. Since the

NOx adsorbers used in these tests were low temperature

(barium based) adsorbers [3,4], only the ESC modes below

50% load were tested. The exhaust temperatures at the ESC

modes above 50% load were too high (> 450 C) for these traps

to adsorb effectively. Due to the potential low temperatureapplication of H2 assisted NOx trap regeneration, a series of

calculated bus road load points (37K lb. GVW) were also

tested. Below bus speeds of 45 mph, the road load points were

at lighter loads with lower exhaust temperatures than the ESC

points. All of the operating points were tested for NO

regeneration using both reformate and diesel fuel injected into

the exhaust as reductants for performance comparison.

Figure 9. Fuel penalty for Bus Road Load, at same NOx

adsorption

Figure 10. NOx adsorption comparison. ESC Modes a

same fuel penalty (reformate vs diesel)

Two comparisons of performance were tested: (1) %NOadsorbed at the same fuel used penalty, and (2) fuel penalty for

the same NOx adsorbed. These two comparisons are made for

both the bus road load points and the tested ESC modes

Figures 8 11 show these results. All of the reformate

regenerations were for a 5 second duration with the fuel

reformer processing 0.8 g/s of fuel. The NOx trap adsorption

time was varied with load to keep the minimum instantaneousNOx adsorption greater than 70%. For the diesel injection

0

50

100

150

200

250

300

350

400

450

500

800 1000 1200 1400 1600 1800 2000

Engine Speed - RPM

Torque

-ft-lb

Road Load ESC 13

45 mph

35 mph

25

15 mph

55 mph

65 mph

0

10

20

30

40

50

60

70

80

90

100

100 150 200 250 300 350 400

Exhaust Temperature - deg C

NoxAdsorbed-%

Diesel Reformate

Upstream DOC Added

0

5

10

15

20

25

30

150 200 250 300 350 400

Exhaust Temperature - degC

Fu

elPenalty-%

FP Diesel FP Reformate

40

50

60

70

80

90

100

280 300 320 340 360 380 400 420 440 460

Exhaust Temperature - deg C

NOxAdsorbed-%

Diesel Reformate


6/7

6

regeneration at constant fuel penalty the total injected fuel

quantity was kept at 4 g and the adsorption time was the same

as the reformate case. For equivalent NOx adsorption with

diesel injection regeneration the injection rate was constant and

the regeneration time varied as well as the adsorption time.

Clearly reformate regeneration of NOx traps is superior at all

the operating points tested both in terms of %NOx adsorbed atthe same fuel penalty and the fuel penalty for equal NOxadsorption.

The reformate regeneration advantage at low exhaust

temperatures is especially obvious in the bus road load

comparisons. Below 180C the diesel injection regeneration

falls from 70% adsorption to zero at 158C. Reformate

regeneration still achieves 90% NOx adsorption at 158C. Atidle conditions with adsorber inlet temperature at 113C

reformate regeneration resulted in 38% NOx adsorption. The

addition of an oxidation catalyst upstream of the NOx trap

increased the NOx adsorption to 75% at 113C with reformate

regeneration. Note that direct injection of diesel fuel into the

exhaust for NOx trap regeneration at these low temperatures is

totally ineffective. In addition other testing with different NO xtrap formulations and simulated reformate has shown

regenerations at exhaust temperatures as low as 80C. The ESC

modes show a 15 to 50% better NOx adsorption at the same fuelpenalty with reformate versus diesel regeneration. At the same

NOx adsorption the ESC modes show a 40 60% fuel penalty

advantage of reformate versus diesel regeneration. These fuel

penalty comparisons include only the fuel used for regeneration

and do not account for required parasitic loads such as

electrical and air pumping power. However at most engine

operating conditions, these parasitic loads account for a fuel

penalty of 0.5%. Only at idle and very light loads are these

parasitics significant. Even then the maximum parasitic fuel

penalty is about 5%.

Figure 11. Fuel penalty for ECS Modes, at same NOx

adsorption

In addition to the fuel penalty advantage, reformate

regenerations of NOx traps produce a dramatic reduction in

hydrocarbon slip as shown in Figure 12. For equivalent NOxadsorption, reformate regenerations reduce hydrocarbon slip

through the adsorbers by 9095% at exhaust temperatures of

275C and above. Below 275C exhaust temperatures the

reduction in HC slip is still 65% or more.

Additional testing completed after this work indicates tha

by optimizing the regeneration time and reformate flow rates

the reformate NOx trap regeneration fuel penalties reported here

can be lowered another 50% from an average of 4% fuepenalty to 2% fuel penalty.

Figure 12. Reduction of hydrocarbon slip when using

reformate vs diesel for NOx trap regeneration.

VEHICLE INSTALLATION AND RESULTSThe H2 assisted NOx trap system installed on a Ford F250

pickup truck, with a 7.3L Powerstroke engine, is illustrated in

Figure 13. This is a single leg with bypass system using 14

liters of NOx adsorber with electrically actuated brake valves to

switch flow legs. The reformer and brake valves are computercontrolled and NOx concentrations are measured with Horiba

NOx sensors. With reformate regeneration of the NOx trap, this

system is capable of 70% NOx adsorption.

Figure 13. F250 hydrogen assisted NOx trap installation

The Gillig transit bus installation is shown in Figures 14 and

15. The bus uses a dual leg system with 21 liter of NO

adsorber per leg. There are separate pneumatically actuated

switching valves for the exhaust flow and reformate flow. A

single Gen-H plasmatron fuel reformer is used. A PC, utilizing

LabView based software, controls the regeneration including

the fuel reformer air-fuel ratio and the switching of the exhaus

brake valves reformate injection NOx trap 14L bypass DOC0

2

4

6

8

10

12

200 250 300 350 400

NOx - ppm

FuelPenalty-%

Diesel Reformate

50

55

60

65

70

75

80

85

90

95

100

150 200 250 300 350 400 450

Exhaust Temperature - Deg C

Reduction-%

Bus Load ESC


7/7

7

and reformate flows on a time based cycle. NOx emissions are

measured with NGK NOx sensors. The engine in this bus is a

MY1994 Cummins C8.3L. Due to the high accessory loads,

engine out NOx concentrations as high as 800-900 ppm are

observed. This system is capable of NOx reductions of 80-90%

(without optimization).

Figure 14. Hydrogen assisted NOx trap installation in a bus

Figure 15 Picture of part of plasmatron diesel reformer system

in the back of the bus

RESULTS SUMMARY

NOx trap regeneration fuel penalty reductions ofroughly 50% versus diesel injection were achieved

Idle operation NOx trap regenerations achieved Hydrocarbon slip during regeneration reduced by 90%

Dual leg NOx trap system installed and operating on atransit bus: 80-90% NOx reduction

Single leg bypass NOx trap system installed andoperating on a Ford F250 truck: 70% NOx reduction

FUTURE EFFORTS Investigate potential advantages of reformate

desulfation of NOx traps: Preliminary tests indicate

400C desulfations are possible.

Test other NOx trap formulations: Improve both highand low temperature performance with reformateregeneration

Develop the ArvinMeritor fuel reformer into a matureproduct

NOMENCLATUREReformate the gas mixture resulting from the partial oxidationof fuel

O/C ratio of oxygen atoms to carbon atoms in the reformer

air-fuel mixture prior to processing

NOx symbol for the variable mixture of NO and NO2 presenin engine exhaust

Fuel Penalty ratio of fuel used for NOx trap regeneration to

the fuel used by the engine over the full adsorption/regeneration

cycle

Hydrocarbon slip hydrocarbon species that flow through the

NOx adsorber during regeneration without reacting

ACKNOWLEDGMENTSThis work was supported by US Department of Energy

Office of Freedom CAR and Vehicle Technology, and by

ArvinMeritor. We thank Dr. S. Diamond for his support and

encouragement of this work. The work conducted a

ArvinMeritor Columbus Technical Center was internally

funded.

REFERENCES

[1] Bromberg L, Cohn DR, Rabinovich A, Heywood

Emissions reductions using hydrogen from plasmatron fue

converters, J, Int. Hydrogen Energy 26, 1115-1121 (2001)

presented at 2000 DEER meeting.

[2] Parks, James, Emerachem, private communication (2002)

[3] Wan, C.Z., Mital, R., Dipshand, J., Low TemperatureRegeneration of NOx Adsorber Systems, Diesel Engine

Exhaust Reduction Workshop 2001, 2001.

[4] Fang, H.L., Huang, S.C., Yu, R.C., Wan, C.Z., Howden, K.

A Fundamental Consideration of NOx Adsorber Technology

for DI Diesel Application, SAE 2002-01-2889, 2002

NOx trap: 21L/leg

fuel reformer box

Documents

BrombergLDEERConf2003